3D printer

Abstract

A 3D printer (1) is described, comprising a print head (6) for extruding a transparent ink strand (8) and a transparent print bed (10) for depositing the extruded ink strand (8). Furthermore, the 3D printer (1) comprises an imaging device (20) that provides a pattern formed from at least straight lines (22), and an optical detection device (14) configured to capture an image (30) of the pattern and the deposited ink strand (8) from an underside (18) of the print bed (10) during normal printing operation of the 3D printer (1). The 3D printer (1) also includes a control unit (12) configured to derive at least one characteristic parameter for printing operation from the image (30) captured by the detection device (14).

Description

[0001] The invention relates to a 3D printing device.

[0002] 3D printers were originally used for prototype production, or at most for small-batch production. The primary goal was to avoid costly processes such as injection molding, which requires expensive injection molds. The simplest form of such "rapid prototyping" is the so-called "Fused Deposition Modeling" (FDM) process, in which thermoplastic material is usually melted and deposited in strands, layer by layer, onto a typically heated build plate, usually in a heated room, using an extruder nozzle. With appropriate process control, a three-dimensional structure with sub-millimeter resolution can be produced by fusing the strands together.The coordinates are usually provided to the 3D printer based on CAD models, which are converted into layer-by-layer paths, known as G-code, using slicer software. Since the development of this "bottom-up" concept, numerous other 3D printing processes have been developed. These sometimes utilize different material classes, such as photocurable resins, which are then solidified stereolithographically to form a component. The use of such manufacturing methods is increasingly being pursued for the production of larger series than small batches, as well as for individual components. Due to the step-by-step manufacturing process, these methods are also collectively referred to as "additive manufacturing."

[0003] Intensive research is also being conducted in the field of regenerative medicine to adapt additive manufacturing processes for the artificial production of tissues or even organs. In so-called biofabrication, living cells are printed together with biocompatible materials to create three-dimensional constructs with defined pore sizes. To enable cell survival and tissue maturation, hydrogel-forming materials are used in particular, which are primarily intended to mimic the natural extracellular matrix of native tissue. For example, biomaterials based on collagen, gelatin, fibrin, hyaluronic acid, alginate, and silk proteins, as well as synthetic polymers such as PEGDA or POx, and various blends and chemically modified versions thereof, are used. The FDM process is frequently used in biofabrication because, unlike, for example,The inkjet process, which is also used for printing cells, allows a wider range of materials to be processed and enables the production of 3D designs in a relatively short time.

[0004] For processing, the materials must first be brought into a liquid, malleable state—e.g., melt, hydrogel, or precursor solution—in order to retain a defined 3D shape after a phase transition. The flow properties (or rheological properties) of the materials to be processed are of significant importance, both for processing and for the dimensional stability of the product. The main characteristic is usually viscosity. It describes the ratio of shear stress to shear rate. Shear stress is the stress that must be applied to deform a fluid with a certain shear rate. "Low viscosity" therefore means that low stresses are required to deform a fluid. These are thin fluids, whereas increasing viscosity (i.e., higher viscosity) indicates a higher viscosity.Higher viscosity (increasing to higher values, often referred to as "high viscosity") describes an increase in "thickness" or "thickness." If there is a linear relationship between shear stress and shear rate, the viscosity is independent of the shear rate and therefore constant – this is called Newtonian flow behavior, which applies, for example, to water and many oils.

[0005] More complex systems, such as polymer melts, high-molecular-weight and concentrated solutions, or multiphase systems like suspensions, often do not exhibit Newtonian behavior, as their viscosity frequently depends on the shear rate or other factors such as time-dependent structural changes in the material. Polymer melts and highly concentrated solutions usually exhibit shear-thinning behavior, meaning their viscosity decreases with increasing shear rate. Physical hydrogels can be destabilized under high shear stress and converted into a liquid state, allowing them to be processed in an extrusion printing process. Such hydrogels, or their precursors (i.e., aqueous solutions), which are loaded with cells and can be converted into a gel immediately after printing (through self-assembly or targeted cross-linking using, for example, ions or irradiation), are referred to below as "bioinks."The print resolution depends on the dimensional stability of the deposited strand, which ideally should spread (i.e., flow apart) as little as possible. How far (or how much) the strand can spread, in turn, depends on the relationship between (temporal) throughput, viscosity, and gelation rate. The challenge in process optimization for bioprinting therefore lies, in part, in systematically capturing the interplay of the involved parameters and optimizing them in a feedback loop. For this purpose, the time-resolved flow behavior of the printed strand must be recorded and documented, quantified, and ideally modeled in relation to specific parameters (pressure, flow rate, deposition rate, temperature, viscosity, etc.). However, capturing the throughput is particularly difficult with relatively low-viscosity materials and / or due to nonlinear flow behavior (e.g.,...Structural viscosity and the like are often difficult to determine solely from process setting parameters (pressure, temperature, and the like).

[0006] The invention is based on the objective of providing an improved 3D printing device.

[0007] This problem is solved according to the invention by a 3D printing device with the features of claim 1. Furthermore, this problem is solved according to the invention by a method with the features of claim 11. Further advantageous and partly inventive developments of the invention are set out in the dependent claims and the following description.

[0008] The 3D printer according to the invention comprises a print head for extruding an ink strand that is at least initially transparent. Furthermore, the 3D printer has a transparent build plate for depositing the extruded ink strand. The 3D printer also includes an imaging device that provides a pattern formed at least from straight, preferably also at least partially parallel, stripes (or lines), as well as an optical detection device configured to capture an image of the pattern and the deposited ink strand from the underside of the build plate during normal printing operation. Finally, the 3D printer has a control unit configured to derive at least one characteristic parameter for printing operation from the image captured by the detection device.

[0009] The term "at least initially transparent" here and in the following refers in particular to the fact that the ink strand, especially the ink used, is transparent after extrusion for at least a certain period of time (e.g., up to 5 or 10 seconds), in particular until solidification (due to cross-linking and / or crystallization) begins or has progressed to a certain degree, and may optionally become opaque thereafter.

[0010] The method according to the invention serves to operate a 3D printer, in particular the one described above. A transparent ink strand is extruded by means of a printhead of the 3D printer (in particular the one mentioned above) and deposited onto a transparent print bed (in particular the one mentioned above). Furthermore, an imaging device (in particular the one mentioned above) is provided, which generates a pattern formed at least from straight, and optionally at least partially parallel, stripes. In addition, an image of the pattern and the deposited ink strand is captured from the underside of the print bed, and at least one characteristic parameter for the printing process is derived from the captured image.

[0011] Preferably, the control unit of the 3D printer is configured to carry out the above-described method, preferably automatically, optionally in interaction with an operator of the 3D printer. Conversely, in optional variants of the method according to the invention, the steps described below as being carried out (or to be carried out) by the control unit are performed accordingly. Thus, the 3D printer and the method exhibit the features and advantages described here and below.

[0012] According to the invention, optical features of the deposited ink strand are detected, and the aforementioned characteristic value is deduced from these features. This is advantageous because it allows for the consideration of features that are a consequence of selected process parameters (e.g., throughput, pressure, feed rate, or the like). However, this detection of the aforementioned characteristic value is decoupled from the process parameters mentioned above. The invention is based on the premise that the image of the pattern detected from below the printing plate is influenced by the transparent ink strand. From this influence, the dimensions of the ink strand, and thus process parameters, particularly their values, can be determined.

[0013] In an optional version, the control unit is configured to use the characteristic parameter to regulate or control the printing process.

[0014] Preferably, the transparent (printer) ink is a bio-ink, i.e., a material that is biologically compatible and specifically suitable and intended for use in or on the human body, e.g., as a tissue replacement, implant, or the like. Such bio-inks or biomaterials are preferably based on collagen, gelatin, fibrin, hyaluronic acid, alginate, and / or silk proteins, or additionally or alternatively on synthetic polymers such as PEGDA or POx, as well as various blends and chemically modified versions thereof.

[0015] In a practical embodiment, the imaging device consists of a plate printed with the pattern. Optionally, this plate can also be transparent and backlit, so that the printed pattern is projected as a kind of shadow onto the printing plate and the ink strand.

[0016] In an alternative embodiment, the imaging device is formed by a screen on which the pattern is displayed.

[0017] In both the case of the printed plate and the screen, however, preferably no projection or the like takes place onto the printing plate and the ink strand. Rather, the pattern is viewed, i.e., captured, by the detection device by looking through the printing plate and the transparent ink strand.

[0018] Advantageously, the imaging device, in particular the plate, is coupled to the print head and is moved along with it.

[0019] In a preferred embodiment, the control unit is configured to infer the (preferably local) width, curvature, and / or height of the deposited ink strand from an optical distortion—e.g., a local rotation, deformation, and / or change in size—of the pattern, particularly caused by the ink strand. This is based on the consideration that a transparent ink strand deposited on the printing plate, due to its spreading—especially under conventional conditions—acts similarly to a (particularly flat-convex) cylindrical lens and thus also locally distorts, in particular rotates, the pattern viewed through the ink strand. The width of the ink strand can, for example, be determined by...The curvature and / or height of the ink strand can be determined relatively easily by identifying the beginning and end of the deformation of a single stripe of the pattern, provided that the angular position of the stripe relative to the direction of travel (especially the printing direction) of the ink strand is known. The curvature and / or height of the ink strand can be deduced, for example, using optical equations. In particular, the path of the light rays through the ink strand, which acts similarly to a cylindrical lens, and through the (at least approximately) plane-parallel printing plate is modeled using formulas from so-called matrix optics; that is, an optical transfer function (or "lens equation") is created in matrix form. Depending on the knowledge of various parameters such as the curvature of the lens, etc., this transfer function can be used to determine the...In particular, the image size (image height, specifically the distance of the "outermost" image point to the optical axis), the object size (object height, again specifically the distance of the "outermost" object point to the optical axis), and the curvature of the lens can be determined. Because the pattern is captured both with and without imaging through the strand, the respective angle of the pattern (in the simplest case, a single strip) relative to the printing direction can be easily read. From the angle ratio (more precisely, the ratio of the two tangents to each other), the ratio of object size to image size can be deduced. Since the image size can be determined (especially on an image sensor of the capture device), the object size can thus be inferred.This allows the curvature of the cylindrical lens formed by the ink strand to be determined using the optical transfer function, particularly given knowledge of the respective refractive indices of the ink strand, the printing plate, and the intervening medium (usually air). From this curvature, the optically detectable width of the ink strand, and the assumption of a parabolic strand profile, the height of the ink strand at its apex and its cross-sectional area can then be determined.

[0020] The stripes of the pattern are preferably positioned obliquely to a main printing direction (i.e., the direction in which the ink strand runs in a straight-line print without skewing, also referred to as the 0° direction). In a particularly advantageous embodiment, the stripes of the pattern are positioned at an angle of at least approximately 45° (i.e., with a deviation of, for example, up to + / - 5°) to the main printing direction. With such an oblique position, any distortion of the pattern or its stripes becomes particularly apparent compared to a perpendicular orientation. With stripes running parallel to the main printing direction, there would be a risk that no stripe would be visible through an ink strand, or that a stripe would be positioned relative to the ink strand in such a way that any distortion would be undetectable or difficult to detect.With an oblique orientation, the probability is particularly high that at least one stripe will cover the ink strand. To achieve a particularly reliable design, it is also conceivable that the pattern is formed from intersecting stripes. A further development of intersecting stripes could be, for example, squares that are spaced apart from each other or directly adjacent to one another and, in particular, have an edge length of 1 to 5 cm, preferably up to 2 cm. Optionally, the stripes, or even just some stripes in a section of the pattern, can be arranged in a star or radial pattern around a common center. For example, these star-shaped stripes are aligned with an angular offset of 5 to 15 degrees to each other. The latter simplifies the detection of distortion or rotation of the stripes in any printing direction.

[0021] Optionally, the stripes can also be interrupted (dotted and / or dashed) and vary in width. This can facilitate the assignment of the optical image (i.e., the picture) of a specific stripe to its "origin" on the imaging device.

[0022] According to a further advantageous embodiment, the control unit is configured to deduce the throughput and / or printing speed (each constituting the aforementioned characteristic quantity) from the (particularly local) width, curvature, and / or height of the deposited ink strand. In particular, the control unit is configured to determine the (local) cross-sectional area of ​​the ink strand based on at least one of the aforementioned characteristics, especially their combination (at least width and curvature), and from this to deduce its volume and volumetric throughput. The local cross-section is known to represent the local throughput. Thus, the (volume) throughput can be determined "inline," i.e., during the process, using a comparatively simple optical observation setup and, if necessary, used for process control or regulation.Given a known printing speed and cross-sectional area (calculated from curvature and strand width), the throughput follows as a process-relevant parameter.

[0023] In an advantageous embodiment, the control unit is also configured to infer the spreading behavior of the ink based on the temporal progression of the width, curvature, and / or height of the deposited ink strand. Specifically, the development of the ink strand is observed in several successive images, and a conclusion is drawn about the spreading behavior, i.e., how much the ink jet spreads out after deposition (especially perpendicular to its printing direction or longitudinal extent). Alternatively, a predefined pattern, i.e., several predefined detection or measurement points along the deposited strand, can be viewed in a single image. The respective measurement points are assigned to different times after the strand is deposited, so that the spreading behavior can be determined in just one image.The spreading behavior, in turn, provides information about process parameters, especially their influence on the 3D printed image (which is affected by the spreading behavior). For example, the degree of pre-gelling or similar can be adjusted if the spreading behavior is outside a target range (too strong or too weak).

[0024] In a preferred embodiment, the control unit is configured to detect, in a calibration mode, a plurality of calibration strands arranged on the printing plate or a carrier to be placed thereon, and which have a known cross-sectional geometry, and to use them to compensate for a measurement deviation. Here and in the following, "measurement deviation" is understood to mean not only a deviation from an "actual" measurement, i.e., for example, the optical determination of the width of the ink strand, but also from a calculation of the curvature and / or height of the ink strand based on optical calculations. In particular, for the known dimensions of the calibration strands, it can be easily verified whether the detection of the width or the calculation leads to results that are at least within a predefined tolerance range.

[0025] Optionally, this calibration mode can be used as a standalone mode in which only the calibration strands are measured. Alternatively, the calibration mode can also be used in parallel with the intended processing, for example, by arranging the calibration strands in a non-printing area on the printing plate where the ink strand is normally deposited. This can also be used, in particular, to compare the deposited ink strand with the calibration strands. In the latter case, it is less a calibration per se and more a kind of interpolation or nearest neighbor method, in which the geometry of the currently considered ink strand is compared with that of the calibration strands.

[0026] Preferably, the calibration strands are made of a transparent material, in particular with a refractive index comparable to that of the (bio-)ink and / or a strand width and / or cross-sectional area comparable to the expected strand widths or cross-sectional areas. In particular, the calibration strands are made of one or more different liquid silicone rubber types. "Comparable" here means, in particular, that the refractive index must lie within a range with a specified tolerance (e.g., + / - 10%) around the value of the corresponding ink size. With regard to strand widths or cross-sectional areas, preferably different calibration strands with widths from 1 mm to 10 mm and variations in strand height from 0.1 mm to 2 mm are used.

[0027] In an optional configuration, the 3D printer also features a radar sensor designed to detect, at least locally (i.e., in particular, as a cross-sectional image), the width, curvature, and / or height of the deposited ink strand during normal printing operation and transmit this information to the control unit. The control unit is then configured, for example, to use this local information as a characteristic parameter when determining throughput and / or printing speed. Specifically, the data (dimensions) detected and transmitted by the radar sensor are used to support or verify the dimensions detected and determined by the optical sensor. The radar sensor is, for example, a UWB ("ultra-wideband") radar sensor.Such a sensor has a comparatively high spatial resolution and the ability to "see through" certain structures (materials). The radar sensor can optionally be directed at the printing plate from below, just like the optical detection device, and must therefore transmit its signal through the plate. Alternatively, however, it is more practical to position the radar sensor above the printing plate in order to measure directly on the deposited ink strand.

[0028] The advantage of the optical detection device (i.e., optical detection itself) lies in the fact that its field of view is generally larger than that of the radar sensor, allowing for a broader area of ​​detection and analysis in a single glance. In comparison, the radar sensor can usually only detect a relatively narrow area (e.g., approximately a line). Nevertheless, the 3D printer with the radar sensor for (at least local) detection of width, curvature, and / or height—in particular, the extrusion geometry—optionally constitutes an independent invention, separate from the presence of the optical detection device. In other words, the 3D printer in this case can include the print head, the (optionally radar-transparent) build plate for depositing the extruded ink strand, as well as the radar sensor and control unit described above.In this configuration, the imaging device and the optical detection device can also be omitted. In this variant, the control unit is specifically designed to deduce the volumetric throughput from the width, curvature, and / or height of the ink jet, as determined by the radar sensor.

[0029] In this context, and in the following, "characteristic" means in particular that the quantity to be determined contains quantitative information about the magnitude of the respective corresponding process or material parameter (e.g., throughput or spreading behavior), such that this information can be unambiguously read from the quantity. However, the quantity to be determined can also be a quantity that is directly or indirectly proportional to the process or material parameter, or that has a non-linear relationship with it, for example, a logarithmic, exponential, or polynomial (i.e., quadratic, cubic, etc.) relationship.

[0030] The conjunction “and / or” is to be understood here and in the following in particular as meaning that the features linked by means of this conjunction can be formed both jointly and as alternatives to each other.

[0031] An embodiment(s) of the invention will be explained in more detail below with reference to a drawing. The drawing shows: Fig. 1. A schematic side view of a 3D printer, Fig. 2. an image of a part of the 3D printer captured by an optical detection device, Fig. 3. In a schematic side view, an optical model for evaluating the image, and Fig. 4 in view according to Fig. 2 an alternative embodiment of the 3D printing device.

[0032] Corresponding parts in all figures are always marked with the same reference symbols.

[0033] In Fig. Figure 1 shows a 3D printer 1. This printer has a gantry frame 2 from which a print unit 4 is suspended. The print unit 4 has a reservoir for the material to be printed (not shown), in this example a "bio-ink" for replicating body tissue. Furthermore, the print unit 4 has a print head, here represented by a print nozzle 6, through which an ink strand 8 is extruded. The 3D printer 1 also has a print bed 10 for depositing the ink strand 8. This print bed is transparent, for example made of glass, or alternatively of polystyrene. The print unit 4 also has a drive mechanism (not shown in detail) by means of which at least the print nozzle 6 (but usually the entire print unit 4) can be moved in a plane parallel to the print bed 10.To control a printing process, the 3D printer 1 has a control unit 12, which in the present embodiment is integrated into the printing unit 4. Alternatively, the control unit 12 can also be a computer that is connected to the printing unit 4 via data transmission.

[0034] To monitor a printing process, the 3D printer 1 also has an optical detection device, specifically a camera 14. This is connected to the control unit 12 via data transmission. The camera 14 is installed such that, in the present embodiment to save installation space, it is arranged below a plane of the printing plate 10, and its field of view is directed from an underside 18 onto the printing plate 10 – more precisely, through it – by means of a mirror 16. Furthermore, the 3D printer 1 has an imaging device 20 that provides a pattern of straight, and in the present embodiment also parallel, stripes 22 (see figure). Fig. 2), which in turn can be viewed (i.e., captured) through the printing plate 10 by means of the camera 14. Specifically, the imaging device 20 is formed by a plate 24 on which the strips 22 are printed. This plate 24 is coupled to the printing unit 4. The printing nozzle 6 protrudes through the plate 24. The plate 24 is therefore moved along with the printing nozzle 6. In an alternative embodiment, the camera 14 can also be positioned so that it "looks" through the printing plate 10 without a mirror, i.e., is aligned vertically with its optical axis.

[0035] The control unit 12 is configured to perform a procedure for operating the 3D printer 1, which is described in more detail below. For this purpose, the control unit 12 continuously captures images of the print bed 10, the ink strand 8 deposited on it, and thus also of the stripes 22, using the camera 14 during a printing process, i.e., while the ink strand 8 is being deposited. Since the deposited ink strand 8 varies in width, exit cross-section (i.e., at the exit of the print nozzle 6), and also its spreading behavior (i.e., how far the ink strand 8 spreads across the print bed 10) depending on process settings such as throughput and printing speed, the stripes are distorted differently by the respective ink strand 8. An example image 30 captured by the camera 14 is shown in Fig. 2 shown.

[0036] The procedure of the control unit 12 is based on the assumption that the deposited and spread ink strand 8 behaves similarly to a cylindrical lens. This is indicated by the fact that the strips 22 are twisted at least in a central longitudinal region of the ink strand 8. The control unit 12 determines the current (and local) width a of the ink strand 8. This can be done, for example, using the optically detectable (longitudinally running) outer edges of the ink strand 8. Furthermore, the control unit 12 determines a curvature of the ink strand 8 using an optical calculation method described in more detail below as an example. Here, the control unit 12 uses the twist of the strips 22, which in the present embodiment run at 45 degrees to a main printing direction 32, and thus also at 45 degrees to the orientation of the ink strand 8.For this purpose, the control unit 12 determines an angle θ for the alignment of the strips 22 (specifically exactly one strip 22) against the main printing direction 32 without deflection by the ink strand 8, as well as an angle θ' for the alignment of a strip 22 imaged by the ink strand 8 and thus deflected (i.e. distorted, in particular twisted).

[0037] Based on Fig. Section 3 now illustrates how the optical calculation method is performed. Specifically, this involves a paraxial approximation using matrix optics. Fig. Figure 3 shows the ink strand 8 in section and forms a cylindrical lens that rests on the printing plate 10. For the beam path originating from an object point, specifically a point that is part of a strip 22 and is arranged at a distance G ("object size") from the optical axis O, the following equation for its exit vector results, assuming that the beam path (here exemplified by a first beam S1) initially runs parallel to the optical axis O up to the ink strand 8. out⇀ on the underside 18 of the printing plate 10: out⇀=B3T3B2T2B1T1hin⇀ where the entry vector of the beam path for beam S1 at inkjet 8, whose vector components represent the distance G and the angle in radians to the optical axis O, T1 strand the translation matrix between the object and the entry into the ink8, B1 the refractive matrix at the entrance to the ink strand 8, T2 the translation matrix in the ink strand 8 to the printing plate 10, B2 the refractive matrix upon passing into the printing plate 10, T3 the translation matrix in the printing plate 10, and B3 describe the refractive matrix as it exits the pressure plate 10.

[0038] Similarly to there⇀ give the components of the vector out⇀ also the distance and the angle in radians to the optical axis of the corresponding ray at the exit point, specifically at the ‘endpoint’ of this ray used for equation (1).

[0039] This can be written out as: hout⇀=[100ng][1D01][100nng][1h01][10(1n−1)ρ1n][1(g−h−D)01][G0] where n g the refractive index of the printing plate 10 (assumed here to be glass), n the refractive index of the ink strand 8, g the distance from the object to the underside 18 of the printing plate 10 (“object distance”), h the local height (or thickness) of the ink strand 8 at the entry point, D the thickness of the printing plate 10, and ρ describes the curvature of the ink strand 8.

[0040] Similarly, the exit vector can be out⇀ The image distance b and the image size Baus at the intersection of the two rays S1 and S2 can also be determined from this.

[0041] From the experimentally accessible relationship tan θtan θ'=GB Given the known object size G, the known object distance g, and the calculable image size B, the curvature ρ can be calculated. Using this and the optically detectable width a of the ink strand 8, the maximum height H can then be determined. h(x)=H(1−(2xa)2)

[0042] The curvature ρ is obtained by differentiating equation (4) twice and taking the absolute value. ρ=8Ha2 which, by rearranging, can again be used to determine the maximum height H, since the curvature ρ can be determined from equation (3).

[0043] Based on the height of the curvature ρ, the cross-sectional area A of the ink strand 8 (and thus also the throughput if the printing speed is known) can now be determined.

[0044] Furthermore, the cross-sectional area A can also be determined from equation (4) by knowing the height of the dog and the width a: A=∫−a2+a2h(x)⋅dx

[0045] This in turn results in: A=23⋅a⋅H

[0046] The product of the cross-sectional area A and the printing speed yields the (volume) throughput of printing ink (in this exemplary embodiment, an alginate-based bio-ink). The throughput is a quantity that is difficult to determine from other process data, especially control variables such as the pressure acting on the ink.

[0047] Basically, the two rays S1 and S2 (or any additional rays passing through the ink strand 8) provide their intersection point "B", i.e., the image size B. However, to solve equation (2), it is also necessary to specify, or rather, to know more precisely, the (local) strand height h. Since the x-coordinate is uniquely determined (calculable) from the geometry and choice of the starting angle of the respective ray S1 or S2 relative to the optical axis, the corresponding local height h can be determined from a value for the maximum height h of width a. Likewise, the curvature ρ contained in equation (2) is given by equation (5). Thus, the maximum height H is the only free parameter of equation (2). The height H will now be expediently varied (iterated) until the calculated ratio of object size G to image size B corresponds to the experimentally accessible tan θtan θ' corresponds, i.e., equation (3) is satisfied.

[0048] The choice of a value for the object size G in equation (2) is basically arbitrary, as long as the value lies in the interval 0 < G < a / 2, preferably in the interval a / 10 < G < a / 4. The first ray S1 is preferably chosen with the starting angle 0 relative to the optical axis (parallel to the optical axis) (see Fig. 3) A suitable starting angle for the second beam S2 relative to the optical axis is determined by the ratio of the selected object size G to the object distance g. The following orders of magnitude can typically be assumed: the object distance g is typically 25 mm (especially + / - 5 mm), the width a of the ink strand 8 is in the range of 0.1 to 1 mm, and the maximum height H is less than the width a. Based on these orders of magnitude, it is clear that the use of the paraxial approximation described above is justified.

[0049] Optionally, the control unit 12 also determines the temporal development of the width of the ink strand 8 based on several consecutive images 30, and optionally also its curvature and / or height. From this, it can be deduced how strongly and for how long the ink tends to spread, or how much the ink strand 8 spreads under given process parameters. This information is important for better assessing and / or controlling the possibilities for the three-dimensional construction of structures.

[0050] To enable the optical measurement of the dimensions and the calculation of the curvature, etc., of the ink strand 8 to be taught and / or verified, several calibration strands 34, specifically two in this case, are arranged on the printing plate 10. These are made of a material with a known optical refractive index (specifically, a refractive index approximating that of the ink), in this embodiment a liquid silicone rubber. In this embodiment, one of the two calibration strands 34 contains a liquid silicone rubber with low spreading and thus a narrow width, while the other, with a higher throughput, contains a liquid silicone rubber with greater spreading and thus a wider width. The width, curvature, and height of the calibration strands 34 are known. This allows verification of the accuracy of the control unit 12's calculations.Alternatively, the control unit can also approximate the dimensions of the ink strand 8 by comparing and interpolating between the known calibration strands 34. The calibration strands 34 are cross-linked ("cured") and can therefore be reused together with the printing plate 10.

[0051] In Fig.Figure 4 shows an image of an alternative imaging device 20. The pattern formed by the strips 22 is configured differently here. The strips 22 form several squares 40 with a side length of 3 cm. As a further embodiment, several strips 22 are arranged in a star-like or radial pattern relative to each other – shown here in one of the squares 40. In the present embodiment, the radials 22 are arranged rotated 15° relative to each other. Such patterns allow any printing direction for the ink strand 8 to be set and still be evaluated with comparatively high reliability according to the scheme described above.

[0052] The subject matter of the invention is not limited to the embodiment described above. Rather, further embodiments of the invention can be derived by a person skilled in the art from the above description. Reference symbol list 1 3D printer 2 portal frames 4 printing units 6 Pressure nozzle 8 Ink strand 10 printing plate 12 Control unit 14 Camera 16 mirrors 18 Underside 20 Imaging device 22 strips 24 plates 30 images 32 Main pressure direction 34 Calibration strand 40 square meters a width θ, θ' angle G distance O pressure plate g distance h height H maximum height Thickness B distance S1 beam S2 beam

Claims

[1] 3D printing device (1), comprising - a print head (6) for extruding a transparent ink strand (8), - a transparent printing plate (10) for depositing the extruded ink strand (8), - an imaging device (20) providing a pattern formed at least from rectilinear stripes (22), - an optical detection device (14) which is designed to detect an image (30) of the pattern and of the deposited ink strand (8) from an underside (18) of the printing plate (10) during the intended printing operation of the 3D printing device (1), and - a control device (12) which is designed to derive at least one variable characteristic of the printing operation from the image (30) captured by the capture device (14). [2] 3D printing apparatus (1) according to claim 1, wherein the imaging device (20) is formed by a plate (24) printed with the pattern. [3] 3D printing apparatus (1) according to claim 1, wherein the imaging device (20) is formed by a screen on which the pattern is displayed. [4] 3D printing device (1) according to one of claims 1 to 3, wherein the control device (12) is configured to infer a width, a curvature and / or a height of the deposited ink strand (8) based on an optical distortion of the pattern in the image (30). [5] 3D printing device (1) according to one of claims 1 to 4, wherein the stripes (22) of the pattern are set at an angle of at least approximately 45° to a main printing direction (32). [6] 3D printing device (1) according to claim 4 or 5, wherein the control device (12) is configured to infer a throughput and / or a printing speed as a characteristic variable from the width as well as the curvature and / or the height of the deposited ink strand (8). [7] 3D printing device (1) according to claim 5 or 6, wherein the control device (12) is configured to infer a spreading behavior of an ink used for the ink strand (8) based on a temporal progression of the width, the curvature and / or the height of the deposited ink strand (8). [8] 3D printing device (1) according to one of claims 1 to 7, wherein the control device (12) is configured to detect, in a calibration mode, a plurality of calibration strands (34) which are arranged on the printing plate (10) or a carrier to be placed thereon and which have a known cross-sectional geometry, and to use them to compensate for measurement deviations. [9] 3D printing device (1) according to claim 8, wherein the calibration strands (34) are formed from a material with a comparable refractive index and / or comparable viscosity as the ink. [10] 3D printing device (1) according to one of claims 1 to 9, comprising a radar sensor which is configured to detect at least locally the width, curvature and / or height of the deposited ink strand (8) during the intended printing operation of the 3D printing device (1) and to supply it to the control device (12) as a characteristic variable for determining the throughput and / or the printing speed. [11] Method for operating a 3D printing device (1), wherein according to the method - a transparent ink strand (8) is extruded by means of a print head (6) of the 3D printing device (1) and deposited on a transparent printing plate (10), - an imaging device (20) providing a pattern formed at least from rectilinear stripes (22) is provided, - an image (30) of the pattern and of the deposited ink strand (8) is captured from an underside (18) of the printing plate (10), and - at least one variable characteristic of the printing operation is derived from the captured image (30).

Images